Studying Child Development: The Importance of Appropriate The Scientific Method Measurement • Treat beliefs about development as hypotheses (educated • Measurements must yield reliable and valid guesses) that need to be tested. results. • Test them using the scientific method: – Choose a question to be answered. – Reliability: Independent measurements of a – Formulate a hypothesis regarding the question. behavior are consistent with each other. – Develop a method for testing the hypothesis. – Validity: The test or experiment measures – Use the data yielded by the method to draw a what it is intended to measure. conclusion regarding the hypothesis.
Reliability Validity
• Reliability: Consistency of measures • Internal validity: Are the effects observed attributable to conditions the researcher – Inter-rater reliability: The amount of intentionally manipulated? agreement between different observers or testers. Agreement should be high. • External validity: Do the conclusions allow – Test–retest reliability: Results should be generalization beyond the particulars of this similar over repeated testing. experiment?
1 Studying Child Development: Studying Child Development: Contexts for Gathering Data Contexts for Gathering Data • Interviews • Naturalistic observation: Study children in their usual – Structured interview: Collecting self-reports from all people being environments—home, play, school. studied. – Clinical interview: In-depth focus on each subject.
Studying Child Development: Correlation and Causation Contexts for Gathering Data • Structured observation: Research based on studying • Variables are attributes that vary across children engaging in designed tasks or situations, individuals and situations. usually in a laboratory. • Behavioral research is concerned with determining if and how many of these variables are related. • Correlational and experimental designs are the two main ways of doing this.
2 Correlations—Five Variations Correlational Designs
• Correlation is the relationship between two variables. • Correlations can be either positive or negative in direction. • The direction and strength of a correlation are indicated by a statistic known as the correlation coefficient. + 0.56
Correlation vs. Causation Experimental Designs • Correlation does not equal • Experiment can prove causation if the participants are causation. similar enough to one another and are tested in same – Correlation does not tell you if one variable causes situation. another. – Random assignment ensures that comparable research – There may be a third variable. subjects submit to same experimental setting. • Finding causation requires an experiment. – Experimental control: One group is tested in the experiment and one group—the control group—is not.
3 Experimental Control Designs for studying development
• Independent variable: The experience that the • Cross-sectional - compare children of different experimental group is exposed to. ages on a given behavior • Longitudinal - follow group of children over • Dependent variable: The behavior that is substantial period of time affected by exposure to the experiment. • Microgenetic - follow group of children closely as behaviors appear.
4